China's Growing Role in World Trade - Clark University NBER Trade Volume (2010) China's...exports....

This PDF is a selection from a published volume from the National Bureau of Economic Research

Volume Title: China's Growing Role in World Trade

Volume Author/Editor: Robert C. Feenstra and Shang-Jin Wei, editors

Volume Publisher: University of Chicago Press

Volume ISBN: 0-226-23971-3

Volume URL: http://www.nber.org/books/feen07-1

Conference Date: August 3-4, 2007

Publication Date: March 2010

Chapter Title: China's Exports and Employment

Chapter Author: Robert C. Feenstra, Chang Hong

Chapter URL: http://www.nber.org/chapters/c10457

Chapter pages in book: (167 - 199)

167

5.1 Introduction

In a series of papers, Dooley, Folkerts- Landau, and Garber (2003, 2004a,b,c, 2005) lay out a vision of a “revived Bretton Woods system” to explain international trade and monetary arrangements today. According to their vision, this system has the following elements:

1. Under the old Bretton Woods system, European countries adopted undervalued exchange rates and capital controls, allowing them to pursue export- led growth. They eventually graduated to fl exible exchange rate and capital mobility, thereby jointly forming a “capital account” region (along with Canada and Latin America).

2. Another group of countries, including Asia and especially China, make up the new periphery and again adopted undervalued exchange rates and capital controls to pursue export- led growth. These countries form a “trade account” region. China, in particular, needs to employ some 200 million persons from the rural area, or 10 to 12 million persons per year in the urban areas, which is facilitated by the infl ow of foreign direct investment (FDI).

3. The United States is at the center, and its budget and current account defi cits have their counterpart in the trade surpluses in Asia. The U.S. current account defi cit is fi nanced through official infl ows from the trade account region and private infl ows from the capital account region.

5China’s Exports and Employment

Robert C. Feenstra and Chang Hong

Robert C. Feenstra is a professor of economics at the University of California, Davis, and a research associate of the National Bureau of Economic Research. Chang Hong is a professor of economics at Clark University.

The authors thank Caroline Freund and Michael Dooley for comments and Zhi Wang for providing the 2000 input- output table used in our calculations as well as other data.

168 Robert C. Feenstra and Chang Hong

1. In fact, Dooley et al. (2004c) motivate the collateral idea by noting that the rest of their theory does not necessarily imply a trade defi cit in the United States as center country; by adding the trade defi cits as collateral, that limitation of the theory is overcome.

2. See The Economist, January 14, 2006, 46–47.

4. The system is sustainable so long as the trade account region continues to fi nance the U.S. trade defi cit and protectionism does not occur. Threats of protectionism are offset by the profi ts earned by foreign investors in the “trade account” region, especially China. Conversely, the trade defi cits run by the United States (or, equivalently, the Treasury bills held by China) are a form of collateral that prevents the Chinese from seizing the assets of foreign fi rms, which would lead the United States to default on its fi nancial obligations.

Some of these various hypotheses are more controversial than others. For example, Wei (2007) objects to the idea that U.S. Treasury bills held by China act as collateral against the Chinese seizing foreign plants, arguing that (a) most FDI in China does not come from the United States but rather from Hong Kong; (b) there is no recent history of China seizing control of foreign fi rms; (c) there is even less history of the United States defaulting on its Trea-sury bill obligations. But this fi nal idea of U.S. Treasury bills acting as col-lateral is not really essential for the rest of the theory, and controversy over it need not detract from the other hypotheses.1 The focus of this paper is on the least controversial of their hypotheses, and that is the idea that expanding exports from China serve to create employment in the urban areas.

Our goal is to quantitatively evaluate this employment hypothesis, that is, to answer the question of how much employment is created by rising Chinese exports. Even this hypothesis is not as straightforward as it might seem. A recent article in the Economist entitled “The Jobless Boom” notes that employment growth has been lower than overall economic growth across various countries of Asia, especially in China, and that this ratio has been falling over time.2 Citing a study by the Asian Development Bank (Felipe and Hasan 2006a,b), the article suggests that the reasons for this weak employ-ment growth has been the shift toward more productive, capital- intensive industries. Dooley, Folkerts- Landau, and Garber (2004a, 4) themselves do not expect the employment growth to come entirely from exports and, in fact, suggest that employment growth of 3 million workers per year in China will come from rising exports.

A logical starting point to determine the employment effect of exports is to look at the calculations from input- output (IO) tables for China, with both the direct and indirect demand for labor from ordinary and processing exports. As reviewed in section 5.2, an increase in ordinary exports of $1,000 (the annual wage in manufacturing in 2000) leads to employment of 0.44 person- years in 2000 and 0.34 person- years in 2002, with much smaller effects from processing exports. But surprisingly, applying these static employment

China’s Exports and Employment 169

coefficients to compute the implied employment gains due to the growth in domestic demand and exports, over 1997 to 2002 and 2000 to 2005, leads to employment gains that vastly exceed the actual employment increase in China. In other words, the static employment coefficients are an unreliable guide to computing the actual employment effects of export growth.

One reason why the static employment effects are unreliable has already been suggested: changes in the industry composition of exports toward more productive industries. This source of aggregate productivity growth is sometimes called the “Denison effect” in the U.S. literature (Nordhaus 1992, 215), as discussed in section 5.3. Shifting toward more productive indus-tries means that the labor needed to produce any given output is reduced. We show in section 5.4 that accounting for the Denison effect reduces the employment impact of exports by about 25 percent from the initial calcula-tions, but we still predict employment gains due to rising exports that are much too high.

Besides the shifting composition of industries, aggregate productivity can rise due to technological progress and capital accumulation. We do not attempt to fully account for this second source of productivity growth, but make a limited attempt by using the growth in wages over time: in our calculations with the IO tables, rising wages means reduced employment growth. We show in section 5.5 that this factor further reduces the employ-ment gains that we can expect from exports to 45 percent of the initial calculation for ordinary exports and 75 percent of the initial calculation for processing exports. These are rules of thumb that can be used to reduce the static employment coefficients for exports.

In section 5.6, we investigate the growth in domestic demand in China over 1997 to 2002, when our data is most complete. Looking fi rst at traded goods (agriculture, mining, and manufacturing) and accounting for the Denison effect, we fi nd that the net employment growth in those sectors due to rising domestic demand is actually negative. That is, the shift toward more produc-tive industries has outstripped the increase in fi nal demand, leading to no net job creation. The only source of employment gains during 1997 to 2002 was in the nontraded sectors, such as construction, and fi nal consumer services like restaurants, health services, education, and so on. Taking into account the same factors as for exports, that is, shifting demand across industries and rising wages, we fi nd that the impact of domestic demand on employment is 75 percent smaller than the initial calculation from the IO table, which gives us another rule of thumb.

Using these rules of thumb we revise the static employment coefficients, and in section 5.7, recalculate the impact of rising exports and domestic demand on labor demand in China. We fi nd the implied employment growth from exports is modest over the 1997 to 2002 period: not more than 2.5 mil-lion jobs added per year. During the 2000 to 2005 period, exports grew much faster, so the employment impact is also higher: exports added as much as


3. We lack specifi c defl ators for components of GDP and trade, and the overall Chinese infl a-tion rate is erratic over this period, including some years of defl ation. Because our trade data is reported in US$, we decide to use constant 2000 US$ to measure all other values, converted with the nominal yuan/dollar rate and using the U.S. Consumer Price Index (CPI).

4. The real GDP fi gures we are using are defl ated by the U.S. CPI, as explained in note 3, and are based on expenditure GDP rather than production GDP. For these reasons, the growth rates differ from those sometimes reported in the press.

7.5 million jobs per year. However, domestic demand led to three times more employment gains than did exports, while productivity growth subtracted the same amount again from employment. This calculation confi rms the suggestion in Dooley, Folkerts- Landau, and Garber (2004a, note 5) that about 30 percent of the employment growth in China will come from rising exports. We conclude in section 5.8 that exports have become increasingly important in stimulating employment in China but that the same gains could be obtained from growth in domestic demand, especially for tradable goods, which has been stagnant until at least 2002 and possibly beyond (Aziz and Cui 2007).

5.2 Employment Gains in China

We begin by reviewing the recent growth in employment, gross domestic product (GDP), and exports in China. Throughout the paper, we focus on the period 1997 to 2005, which gives us two overlapping fi ve- year intervals to work with: 1997 to 2002 and 2000 to 2005. Despite the relatively short span of years and closeness of these two intervals, we will fi nd substantial changes in the Chinese economy over this time.

In table 5.1, we list total employment, broken down by urban and rural, as well as GDP and its components during these years. Total employment has grown by 7.5 to 8 million workers per year over this period, while urban employment has grown slightly faster: 8 to 8.5 million workers, as there was some migration out of the countryside. Real GDP and its components, as well as all trade data, is measured in constant 2000 US$.3 Real GDP growth doubled from 5.7 percent per year over 1997 to 2002 to 11.0 percent in 2000 to 2005.4 Notice that the growth of C � G is much less in the 2000 to 2005 period than is the growth in investment, indicating that an increasing share of domestic demand is for construction projects and other investments.

In table 5.2, we provide the data on Chinese ordinary and processing trade, again in constant 2000 US$. Both exports and imports grew by more than 20 percent per year over the 2000 to 2005 period, which greatly outstripped their prior growth: the boom in Chinese trade is really a feature of the twenty- fi rst century. Note that the trade balance listed in the fi nal column of table 5.2 does not match the values for (X – M) given in the fi nal column of table 5.1 because (X – M) includes both goods and services as used in GDP accounts, whereas the trade balance in table 5.2 is just for merchandise trade.


A logical starting point to determine the impacts of export growth on employment are the studies by Chen et al. (2004), using a 1995 IO table for China, and Lau et al. (2006b,c), using a 2002 IO table; both of these esti-mates are summarized in Chen et al. (2008). From the 1995 table, Chen et al. (2004) estimate that $1,000 of ordinary exports from China leads to 0.70 person- years of employment, and $1,000 of processing exports leads to 0.06 person- years, or roughly one- tenth as much as for ordinary exports. Those estimates are shown in table 5.3 and have been falling over time. Using the 2002 IO table, Lau et al. (2006b,c) estimate that $1,000 of ordinary exports from China leads to 0.36 person- years of employment (which is one- half as much as they found for 1995), and $1,000 of processing exports leads to 0.11 person- years (which is twice the estimate for 1995), so processing exports lead to about three- tenths the employment of ordinary exports.

We will refer to these employment estimates computed from the IO tables

Table 5.1 China’s employment and gross domestic product (GDP)

Millions of persons U.S.$ billions, 2000

Year Employment Urban Rural GDP C � G Investment X – M

1997 698 208 490 1,057 623 388 462000 721 232 489 1,193 743 421 292002 737 248 490 1,392 829 527 362005 758 273 485 2,009 1,043 856 110

Growth (million per year) Growth rate (% per year, compound)

Employment Urban Rural GDP C � G Investment X – M

1997–2002 7.8 8.0 –0.2 5.7 5.9 6.3 –4.92000–2005 7.5 8.4 –0.9 11.0 7.0 15.3 30.7

Source: China Statistical Yearbook, various years.

Table 5.2 China’s ordinary and processing trade

Year Ordinary exports

Processing exports

Total exports

Ordinary imports

Processing imports

Total imports

Trade balance

1997 89 107 196 77 75 152 442000 112 138 249 133 93 225 242002 139 172 312 166 117 283 292005 305 367 672 340 242 582 90

Growth rate (% per year, compound)1997–2002 9.3 10.0 9.7 16.5 9.2 13.1 –7.92000–2005 22.2 21.7 21.9 20.8 21.2 20.9 30.1

Source: China customs trade data.


as “static employment coefficients” because they each refer to a single year. The change in these static employment coefficients can be due to either of the factors we identifi ed in the introduction: shifting composition of exports across industries and technological progress. We will attempt to measure the importance of each of these but fi rst need to replicate the results of Lau and his coauthors for one year. Using the IO table for 2000, we fi nd that $1,000 of ordinary exports from China leads to 0.44 person- years of employment, and $1,000 of processing exports leads to 0.13 person- years. So again, processing exports leads to about three- tenths the employment of ordinary exports. Our estimates for 2000 are also shown in table 5.3 and fall neatly in between the estimates of Chen et al. (2004) and Lau et al. (2006b,c), giving us some confi dence that our employment estimates are consistent with theirs.

The methodology we have used to obtain the static employment coefficients from the 2000 IO table is discussed in the appendix and is briefl y summa-rized as follows. Denote the sources of demand by j � D, O, P for domestic demand, ordinary exports, and processing exports, respectively. Then the portion of value added going to labor from $1 demand of type j in sector i is B j

Lit, which is computed from the IO table as the sum of direct plus indi-rect payments to labor. Our calculations are only for 2000, which we denote t � 0, but the same calculations are made by Chen et al. (2004) and Lau et al. (2006b,c) for 1995 and 2002. Having obtained these coefficients B j

Lit for each sector, these are averaged across sectors:

(1) B�DLt �

∑iDitBDLit

��∑iDit

, and B�jLt �

∑iXjitB

jLit

��∑iX

jit

, for j � O, P,

where Dit denotes domestic demand in sector j, while XitO denotes ordinary

exports, and XitP denotes processing exports.

Notice the averaged terms B� jLt refer to the portion of value added going

Table 5.3 Static employment coefficients (implied employment increase per $1,000 of exports or domestic demand in person- years)

Source Ordinary exports

Processing exports

Domestic demand

Chen et al. from 1995 IOa 0.703 0.057 n.a.Our estimates 2000 IOb 0.444 0.130 0.562

Lau et al. from 2002 IOc 0.363 0.111 0.492

Note: IO � input- output; n.a. � not available.aChen et al. (2004, tables 7, 8) and also Chen et al. (2008, table 1).bAuthor estimates for 2000 IO table, as described in the appendix.cLau et al. (2006c, table 4) for ordinary and processing exports and also Chen et al. (2008, table 1), with domestic demand coefficient computed as explained in the text.


to labor. To convert this into employment, we need to divide by a wage. For the 2000 IO table, we have used the average 2000 wage, which was $842 per year. So the static employment coefficients shown in table 5.3 for 2000 are obtained as:

(2) C�jL0 � B�j

L0/$842, for j � D, O, P,

We are unsure what wages were used by Chen et al. (2004) and Lau et al. (2006b,c) for 1995 and 2002, but the calculation is presumably analogous to that in equation (2), which we will write in other years as:

(2�) C�jLt � B�j

Lt/Wt.

In table 5.3 we also show the static employment coefficient for domestic demand, which equals C � I � G. For 2000, we have computed the domestic coefficients as in equations (1) and (2), for j � D. For 2002, we choose C�j

Lt so that the implied employment from domestic demand plus exports just equals the actual employment in each year. That is, we have chosen C�D

Lt so that:

(3) C�DLtDt � C�O

LtXtO � C�P

LtXtP � Lt,

where Lt is employment in year t. Notice that this full- employment condi-tion also holds in 2000 by construction of the static employment coefficients from the IO table.

Despite the fact that the static employment coefficients are obtained for a single year, there is a strong temptation to apply them over time, that is, to use these coefficients to predict the future course of employment due to export growth. There are potentially large errors associated with that procedure, however. To see this point theoretically, take the difference of equation (3) over a fi ve- year period. After some simplifi cation, we obtain the equation:

(4) �Lt � �Dt 1�2

(C�DLt � C�D

Lt�5) � �XtO

1�2

(C�OLt � C�O

Lt�5)

� �XtP

1�2

(C�PLt � C�P

Lt�5) � �C�DLt

1�2

(Dt � Dt�5)

� �C�OLt

1�2

(XtO � XO

t�5) � �C�PLt

1�2

(XtP � XP

t�5),

where �Dt � Dt – Dt–5 is the change over a fi ve- year interval and likewise for every other variable. On the fi rst line of equation (4), we have the change in domestic demand and exports times the average employment coefficients, and on the second line we have the change in the employment coefficients times the average demand. Generally, the employment coefficients are falling over time, as can be seen by comparing the rows of table 5.3. It follows that the last three terms of equation (4) is negative and potentially quite large: the fall in each employment coefficient is multiplied by the average level


5. The rise in domestic demand of $411.5 billion in table 5.4 is taken from the IO tables for 1997 and 2002 and exceeds the rise in domestic demand of $345.4, taken from C � I � G in the national accounts, table 5.1. We use domestic demand from the IO tables for consistency with later calculations. If instead we use the national accounts fi gure to predict employment gains in table 5.4, then we obtain 225 million workers over 1997 to 2002, which is somewhat less than what we report in table 5.4.

of demand and not just its growth. Because these terms are negative and potentially large, it follows that the fi rst three terms on the right are poten-tially much larger than the actual increase in employment.

This theoretical result is confi rmed in table 5.4, where we take the static employment coefficients and apply them to the change in exports over the two fi ve- years periods, 1997 to 2002 and 2000 to 2005. In the fi rst row of table 5.4, for example, we use our estimates of the C j

i0 employment coefficient from the 2000 IO table, as shown in table 5.3, and multiply each of the employ-ment coefficients by the real change in domestic demand, ordinary exports, and processing exports over 1997 to 2002. That is, we compute:

(5) Prediction 1 � �DtC�DL0 � �X t

OC�OL0 � �Xt

PC�PL0,

which is similar to the fi rst three terms of equation (4). From domestic demand, we predict an employment increase of 216 million persons, and for ordinary processing exports, we predict an employment increase of 31 million persons.5 Summing over these, we obtain nearly 250 million work-ers, as compared to an actual employment increase of only 39 million! We see that simply multiplying the real changes in demand and exports by the employment coefficients, as in equation (5), massively overstates the true change in employment.

The situation is even worse over the 2000 to 2005 period, where now we use the static employment coefficients of Lau et al. (2006c) from the 2002 IO table. Again, we multiply the employment coefficients by the real change in domestic demand and exports, as in equation (5), and predict an increase in employment in China of 550 million workers, as compared to the actual increase of only 37 million! Thus, the predicted employment impact vastly exceeds the actual employment increase. The difference between the pre-dicted and actual employment increases is due to fall in the employment coefficients, as shown by the fi nal terms of equation (4).

We conclude from these calculations that the static employment coeffi-cients, times the changes in demand, do not provide reliable estimates of the actual employment gains in China. Reasons for this have already been suggested: the static employment coefficients do not take into account the changing industry composition of domestic demand and exports, and the coefficients can also fall due to technological progress and capital accumula-tion. We now examine each of these explanations in turn.

Tab

le 5

.4

Impl

ied

Chi

nese

em

ploy

men

t fro

m d

omes

tic

dem

and

and

expo

rts

(usi

ng s

tati

c em

ploy

men

t coe

ffici

ents

)

Gro

wth

in d

eman

d (U

.S.$

bi

llion

s, 2

000)

Impl

ied

empl

oym

ent i

ncre

asea

(mill

ions

of

pers

ons)

Act

ual e

mpl

oym

ent

incr

ease

b (m

illio

ns

of p

erso

ns)

Sour

ce

Per

iod

D

omes

tic

dem

and

E

xpor

ts

Dom

esti

c de

man

d

Exp

orts

To

tal

Aut

hor

esti

mat

es fr

om 2

000

IO19

97–2

002

411.

511

5.5

231.

130

.826

1.8

39.2

Lau

et a

l. (2

006c

) fro

m 2

002

IO

2000

–200

5

735.

0

422.

7

361.

6

95.6

45

7.2

37

.4

Not

e: I

O �

inpu

t- ou

tput

.a U

ses

the

stat

ic e

mpl

oym

ent c

oeffi

cien

ts fr

om ta

ble

5.3

and

mul

tipl

ies

thes

e by

the

real

gro

wth

in d

omes

tic

dem

and,

ord

inar

y ex

port

s, a

nd p

roce

ssin

g ex

port

s.b A

ctua

l em

ploy

men

t inc

reas

e co

mes

from

the

Chi

na S

tati

stic

al Y

earb

ook

of e

ach

year

(see

tabl

e 5.

1).


6. This industry also had by far the greatest increase in real exports over 2000 to 2005, exceed-ing $160 billion, though the majority of those sales were for processing exports.

5.3 Shifting Composition of Exports and Domestic Demand

The static employment coefficients computed from the IO table refer to the employment impact of an additional $1,000 in average exports or domes-tic demand, that is, using the same composition of output that occurred in the year of the IO table, as shown by taking the averages in equation (1). But that is not a good guide for the effects of an actual change in demand because with shifting comparative advantage, export growth may be in industries different from in the past. In addition, for domestic demand, the growth in China in recent years has been especially strong in investment (as shown in table 5.1), especially construction, which differs in its labor requirements from other industries.

The growth in exports is shown in fi gures 5.1 and 5.2, where we graph the percentage increase over 2000 to 2005 in total and ordinary exports, respectively, and industry wages in 2000. Regardless of whether we use total or ordinary exports, the industry with the greatest percentage increase in exports was electronic and telecommunications equipment, and that indus-try also had the highest wage in 2000.6 Overall, there is a positive correlation between the percentage growth in exports, and the real wage in 2000, with food products and tobacco appearing as an outlier (and a relatively small export industry). The fact that the percentage increase in exports differs substantially across industries, meaning that the use of “average” exports as

Fig. 5.1 Growth in total exports, 2000–2005, and industry wages, 2000


in Chen et al. (2004) and Lau et al. (2006b,c) will lead to inaccurate results. Instead, we want to use the “marginal” exports, that is, the actual increase in exports that occurred in each industry over the fi ve- year period.

In theoretical terms, we want to compare the results of using aggregate employment coefficients, as shown in table 5.3, with using disaggregate sector- level coefficients. To obtain the disaggregate results, write the full- employment condition equation (1) alternatively as:

(6) ∑i

CDLitDit � CO

LitXitO � CP

LitXitP � ∑

i

Lit,

where CDLit, C

OLit, and CP

Lit are the disaggregate employment coefficients by IO sectors and likewise for domestic demand Dit, ordinary exports Xit

O, and processing exports Xit

P. Taking the difference of equation (6) over a fi ve- year interval, we obtain:

(7) �Lt � ∑i

��Dit1�2

(CDLit � CD

Lit�5) � �XitO

1�2

(COLit � CO

Lit�5)

� � itXP

1�2

(CPLit � CP

Lit�5)� � ∑

i

��CDLit

1�2

(Dit � Dit�5) � �COLit

1�2

(XitO � XO

it�5)

� �CPLit

1�2

(XitP � XP

it�5)�.

Fig. 5.2 Growth in ordinary exports, 2000–2005, and industry wages, 2000


By using the sectoral data in the 2000 IO table, we can make an alternative prediction of the employment gains from the fi rst two lines of equation (7):

(8) Prediction 2 � ∑i

(�DitCDLi0 � �Xit

OCOLi0 � �Xit

PCPLi0),

where we are using employment coefficients from the year 2000 table in place of the average employment coefficients that appear in equation (5). Note that these are obtained from the 2000 IO table by dividing B j

Li0 by the wage in each sector:

(9) Prediction 2 uses: C jLi0 �

B jLi0

�Wi0

, for j � D, O, P.

Comparing the new prediction obtained from the disaggregate coefficients in equation (8) with that from the aggregate coefficients in equation (5), because �Dt � �i�Dit and �Xj

t� �i�Xjit, we obtain:

(10) Prediction 2 � Prediction 1

� ∑i

[�Dit (CDLi0 � C�D

L0) � �XitO(CO

Li0 � C�OL0) � �Xit

P(CPLi0 � C�P

L0)].

If there is a negative correlation between the growth in demand and the employment coefficients in each sector, as we would expect if growth in output occurs in the more efficient sectors, then equation (10) is negative, and our second prediction of employment growth is less than the fi rst. This reduction in employment gains comes from shifts toward more produc-tive industries and is an example of what Nordhaus (1992, 215) calls the “Denison effect.” Nordhaus refers to the work of Edward Denison (1967, 1980), who demonstrated that if resources shift from low- productivity to high- productivity industries, like from agriculture to manufacturing, then the economy would show aggregate productivity growth even if sectoral productivity growth was zero in both sectors. The aggregate productivity growth is due to a “reallocation effect” across industries. The fl ip side of this aggregate productivity growth is that the labor needed to produce any given output is reduced, as we are showing in equation (10).

Another interpretation of the calculation in equation (8) can be obtained by taking the averages:

(11) CDL0 �

∑ i�DitCDLi0

��∑ i�Dit

, and C jL0 �

∑i�XjitC

jLi0

��∑i�Xj

it

, for j � O, P.

Notice that equation (11) is an average of the sectoral employment coefficients Cj

Li0 in 2000, but using the change in domestic demand and exports as weights, rather than their average levels as in equations (1) and (2). Again, because �Dt � �i�Dit and �Xj

t� �i�Xjit, it is immediate that prediction 2 in equation

(8) can be alternatively written as:


7. If instead of using the industry wages in prediction 2, as in equation (9), we instead con-tinued to use the overall average wage of $842 in 2000, then the predicted employment impact of exports is reduced by 15 percent as compared with the fi rst prediction. That reduction comes from using the disaggregate calculation as in equation (8), but with the average wage of $842 in equation (9). The additional 10 percent reduction for prediction 2 is obtained by using the industry wages, as in equation (9).

(12) Prediction 2 � �DtCDL0 � �Xt

OCOL0 � �Xt

PCPL0,

which is the change in demand times the revised employment coefficients. From equations (11) and (12), we can see our second prediction of the rise in employment uses actual or “marginal” increase in exports and domestic demand, rather than the “averages” used in equations (1) and (2) and equa-tion (5).

In the following sections, we implement this second prediction, as well as a third variant, using the 2000 IO table. In sections 5.4 and 5.5, we focus on the growth of exports, over 1997 to 2002 and 2000 to 2005, and in section 5.6 discuss the growth in domestic demand, in which case we do not have disag-gregate data for 2005, so we are restricted to investigating 1997 to 2002.

5.4 Growth of Exports, 1997 to 2002 and 2000 to 2005

In table 5.5, we report the employment gains over 1997 to 2002 and 2000 to 2005 using the disaggregate increase in exports over these two periods (prediction 2a). In the former period, 1997 to 2002, the employment growth is 22.7 million persons, rather than 30.8 million from table 5.1. So the shift toward more productive industries reduces the employment growth by 25 percent (or 17 percent for ordinary exports and 52 percent for processing exports). A similar decline is seen over 2000 to 2005, when using the actual rather than the average increase in exports reduces employment growth from 115.4 million (prediction 1) to 86.1 million (prediction 2a), again a decline of 25 percent.7 We conclude that the employment gain from increased exports is reduced once we account for the industry composition of exports, as sug-gested by Felipe and Hasan (2006a,b).

The adjustments we have made for prediction 2a can be extended in two directions: we have the data to take into account the provincial compositions of exports, along with provincial wages by industry; or to account for the differing wages paid by types of fi rm ownership (state- owned, collective, or private) and the exports by fi rm ownership and industry as well as wages by fi rm ownership and industry. To the extent that exports are shifting to more productive provinces (e.g., coastal) or fi rms (e.g., private), the estimated employment gains are reduced.

It should be noted that the maintained assumption in these calculations is that the national IO table for 2000 applies equally well across provinces and across types of fi rm ownership. We have only very limited data that could be

Tab

le 5

.5

Impl

ied

incr

ease

in C

hine

se e

mpl

oym

ent f

rom

exp

orts

(usi

ng in

put-

outp

ut ta

ble

in 2

000,

and

indu

stry

wag

es in

200

0)

Impl

ied

empl

oym

ent i

ncre

ase

(mill

ions

of

pers

ons)

Per

cent

age

redu

ctio

n fr

om

Pre

dict

ion

1

Aut

hor

esti

mat

es u

sing

200

0 in

put-

outp

ut ta

ble

P

erio

d

Ord

inar

y ex

port

P

roce

ssin

g ex

port

To

tal e

xpor

t

Ord

inar

y ex

port

P

roce

ssin

g ex

port

Pre

dict

ion

1, fr

om ta

ble

5.4:

usi

ng a

vera

ge

expo

rts,

and

ave

rage

wag

es in

200

019

97–2

002

22.3

8.5

30.

820

00–2

005

85.7

29.7

115.

4P

redi

ctio

n 2a

: usi

ng in

dust

ry e

xpor

ts, a

nd

indu

stry

wag

es in

200

019

97–2

002

18.6

4.1

22.

717

5220

00–2

005

69.4

16.7

86.

119

44P

redi

ctio

n 2b

: usi

ng in

dust

ry- p

rovi

nce

expo

rts,

and

indu

stry

- pro

vinc

e w

ages

in

2000

1997

–200

217

.0 3

.5 2

0.6

2459

2000

–200

563

.214

.4 7

7.5

2652

Pre

dict

ion

2c: u

sing

fi rm

- ow

ners

hip

expo

rts,

an

d fi r

m- o

wne

rshi

p w

ages

in 2

000

19

97–2

002

17.2

3.9

21.

023

5420

00–2

005

59

.1

15.1

7

4.1

31

49

Sou

rce:

Aut

hors

’ cal

cula

tion

s as

exp

lain

ed in

the

text

.


used to test this assumption. To the extent possible, we applied the methods of Bernstein and Weinstein (2002) and found that the 2000 IO table appears to hold reasonably well across provinces except for Guangdong (where labor compensation was higher than predicted from the national IO table). Because Guangdong was the only outlier, and because our ability to construct an alternative IO table for Guangdong is extremely limited, we continued to apply the national table across all provinces and types of fi rm ownership.

Focusing fi rst on the provincial effects (prediction 2b), accounting for the shift in exports by industry and province further reduces the employment impact of increased exports, to 20.6 million persons over 1997 to 2002, or one- third less than the initial calculation. For 2000 to 2005, the implied increase in employment is 77.5 million persons, which is also one- third less than the initial calculation. The employment effects that are obtained when we instead take into account the shift in exports by industry and fi rm own-ership (prediction 2c) are similar to those that take into account provincial effects: the predicted employment gains are reduced by about one- third from the initial calculations. The data we have available do not allow us to take into account both of these effects at the same time. In any case, for 2000 to 2005, the implied increase in employment is still much larger than the actual increase of 37 million, which calls for an explanation.

5.5 Increase in Wages due to Productivity Gains

A fi nal limitation of the static employment coefficients computed from the IO table, and also a limitation of our results reported in table 5.5, is that we have assumed that wages are constant over time. That is, we are using wages in 2000: either at the overall wage in equation (2) or the industry wage in equation (9). But, of course, real wages will rise over time due to both productivity gains and capital accumulation. With rising wages, any implied increase in value added and payments to labor will correspond to a smaller increase in employment.

For our next calculation, we divide the direct plus indirect payments to labor from the 2000 IO table by the real 1997 and 2002 wages, respectively, when estimate labor demand in each year. That is, we obtain the employment coefficients in each year as:

(13) C jLit �

B jLi0

�Wit

and C jLit�5 �

B jLi0

�Wit�5

, for j � D, O, P.

Then our third prediction of the employment gains for rising demand is:

(14) Prediction 3 � ∑i

(DitCDLit � Dit�5C

DLit�5 � Xit

OCOLit � XO

it�5COLit�5

� �XitPCP

Lit � �XPit�5C

PLit�5).

Note that if instead of the estimates in equation (13), we had used the true employment coefficients C j

Lit obtained from the IO table in each year, then


8. Imports are treated entirely as intermediate inputs in the IO table, so need not be deducted from C � I � G.

equation (14) would be an exact prediction of the change in employment: there would be no error involved. So the difference between the third predic-tion, which uses the industry wages in each year, and the actual changes in employment occurs because (a) we are using wages in equation (13) that do not differ between domestic and export production, and (b) we are still using coefficients B j

Li0 from 2000 rather than allowing these coefficients to change over time. In brief, we still do not accurately predict employment changes with equation (14) because we are not allowing the IO table to change over time, and our wage data is not detailed enough. Still, we fi nd that this third prediction is a further improvement over our earlier calculations.

In table 5.6, we show how the implied employment effects are further reduced when we allow for the actual increase in wages over 1997 to 2002 or 2000 to 2005. For 1997 to 2002, we fi nd that the employment gains due to ordinary exports range from 5 to 10 million (predictions 3a, 3b, and 3c), which are reduced by 55 percent or more as compared to the initial calcula-tion. For processing exports, the implied employment effects range from –1.4 to 1.7 million, a reduction of at least 80 percent from the initial calculation. Over this period, most of the increase in exports over these years can be explained by the shift in workers toward more efficient industries, fi rms, and provinces, so the employment gain is very modest. Over 2000 to 2005, we also fi nd that the employment gains due to increased ordinary exports are reduced by 55 percent from our initial calculation, while the employment gains due to processing exports are reduced by about 75 percent.

To sum up, our calculations have reduced the employment impact of increased exports by more than one- half of the initial calculation for ordinary exports, and at least three- quarters for processing exports. Are these results in table 5.6 believable? The smaller employment gains indicate an efficient reallocation of resources, which is plausible. We note that these efficiency gains come from reallocations across many industries (as well as province and fi rm ownership) and do not simply refl ect a rural- urban migration. Indeed, agriculture and manufacturing industries tend to rise or fall together in our calculations: allowing for rising wages over time, we fi nd that the increase in exports is associated with rising employment in both agriculture and the sum of all manufacturing industries. So the net changes in implied employment reported in tables 5.5 and 5.6 would be similar if we omitted agriculture and reported instead the changes in manufacturing employment due to exports.

5.6 Shifting Composition of Domestic Demand, 1997 to 2002

To measure domestic demand, we rely on the sum of C � I � G by indus-try from the IO tables, which we have for the years 1997, 2000, and 2002, but not for 2005.8 So to evaluate the change in employment due to domestic

Tab

le 5

.6

Impl

ied

Chi

nese

em

ploy

men

t fro

m e

xpor

ts (u

sing

inpu

t- ou

tput

tabl

e in

200

0, a

nd in

dust

ry w

ages

by

year

)

Impl

ied

empl

oym

ent i

ncre

ase

(mill

ions

of

pers

ons)

Per

cent

age

redu

ctio

n fr

om

Pre

dict

ion

1

Est

imat

es u

sing

200

0 in

put-

outp

ut ta

ble

P

erio

d

Ord

inar

y ex

port

P

roce

ssin

g ex

port

To

tal e

xpor

t

Ord

inar

y ex

port

P

roce

ssin

g ex

port

Pre

dict

ion

2a, t

able

5.5

: usi

ng fi

rm-

owne

rshi

p ex

port

s, a

nd fi

rm- o

wne

rshi

p w

ages

in 2

000

1997

–200

217

.2 3

.921

.023

54

2000

–200

559

.115

.174

.131

49

Pre

dict

ion

3a: u

sing

indu

stry

exp

orts

, and

in

dust

ry w

ages

by

year

1997

–200

2 4

.7–1

.4 3

.379

117

2000

–200

537

.9 7

.845

.756

74

Pre

dict

ion

3b: u

sing

indu

stry

- pro

vinc

e ex

port

s, a

nd in

dust

ry- p

rovi

nce

wag

es b

y ye

ar

1997

–200

2 5

.7 0

.1 5

.874

99

2000

–200

536

.5 7

.844

.457

74

Pre

dict

ion

3c: u

sing

fi rm

- ow

ners

hip

expo

rts,

an

d fi r

m- o

wne

rshi

p w

ages

by

year

19

97–2

002

10.3

1.7

12.0

54 8

0

2000

–200

5

38.3

8

.1

46.4

55

7

3

Sou

rce:

Aut

hors

cal

cula

tion

s as

exp

lain

ed in

the

text

.


9. Tradable goods are defi ned as sectors 1–22 of the 2000 IO table, and nontradable goods as sectors 23–40.

demand, we are restricted to the fi ve- year period 1997 to 2002 and will not be able to report any results for 2000 to 2005. Furthermore, domestic demand is not broken down by province nor by the type of fi rm ownership. So the calculations for domestic demand will only be broken down by industry over 1997 to 2002.

The implied employment increase due to the growth in domestic demand are reported in table 5.7, where we distinguish domestic demand for trad-able goods (all manufacturing plus mining and agriculture) and nontraded goods (all utilities and services, including construction).9 That is, we have recomputed the employment coefficients shown in equations (1) and (2) and equations (10) and (11) for domestic demand by separating traded from nontraded goods. Traded goods are shown in panel A of table 7. We fi nd that domestic demand for tradable goods has risen by a very modest amount in real terms over 1997 to 2002, $24 billion, shown in the fi rst column. Multi-plying that increase in demand by the static employment coefficient of 0.525, we obtain a modest rise in employment of 12.7 million persons, as shown in the third column (prediction 1).

However, if instead we use the actual change in demand rather than its “average” change, then fall in demand would actually lead to reduced employment of 9.9 million workers when holding wages fi xed at their 2000 levels (prediction 2). Allowing for the growth of wages between 1997 and 2002, the implied fall in employment is even higher, 49.8 million workers, due to the fall in domestic demand (prediction 3). Only a very small amount, 3.3 million workers, is made up by the increase in demand due to rising exports, so the net change in employment due to domestic demand plus exports is a fall of some 47 million jobs.

Because employment actually increased by 39 million jobs over 1997 to 2002, the gap must be made up by the nontraded sector, which is confi rmed in the next row of table 5.7. An initial calculation using a static employment coefficient gives a rise in employment of 203 million (prediction 1, panel B). Use the actual change in demand rather than its “average” change, then the employment increase becomes 166 million workers when holding wages fi xed at their 2000 levels (prediction 2). Allowing wages to rise over 1997 to 2002, the employment gain in nontradable goods is 111 million workers (prediction 3). That is an enormous rise in employment due to domestic demand, which far exceeds any of our calculations for exports. The sector with the largest increase in domestic demand is construction, which accounts for at least half of the overall rise in employment. Employment gains are also shown in fi nal consumer services like real estate, restaurants, health services, education, and so on.

The changes in domestic demand for tradable and nontradable goods

Tab

le 5

.7

Impl

ied

Chi

nese

em

ploy

men

t fro

m d

omes

tic

dem

and

and

expo

rts,

199

7–20

02 (u

sing

inpu

t- ou

tput

tabl

e in

200

0 an

d in

dust

ry w

ages

, 200

0 or

by

yea

r)

Gro

wth

in d

eman

d (U

.S.$

bi

llion

s, 2

000)

Impl

ied

empl

oym

ent

incr

ease

(mill

ions

of

pers

ons)

Per

cent

age

redu

ctio

n fr

om

Pre

dict

ion

1

Aut

hor

esti

mat

es u

sing

200

0 in

put-

outp

ut ta

ble

D

omes

tic

dem

and

To

tal e

xpor

t

Dom

esti

c de

man

d

Tota

l exp

ort

D

omes

tic

dem

and

To

tal e

xpor

t

A. T

rade

d go

ods

Pre

dict

ion

1: u

sing

ave

rage

dem

and,

ave

rage

wag

es in

200

0 2

4.1

115.

512

.730

.8P

redi

ctio

n 2:

usi

ng in

dust

ry d

eman

d, in

dust

ry w

ages

in 2

000

24.

111

5.5

–9.9

22.7

–178

26.3

Pre

dict

ion

3: u

sing

indu

stry

dem

and,

indu

stry

wag

es b

y ye

ar 2

4.1

115.

5–4

9.8

3.3

–492

89.3

B. N

ontr

aded

goo

dsP

redi

ctio

n 1:

usi

ng a

vera

ge d

eman

d, a

vera

ge w

ages

in 2

000

387.

40

203.

30

Pre

dict

ion

2: u

sing

indu

stry

dem

and,

indu

stry

wag

es in

200

038

7.4

016

5.7

018

.5n.

a.P

redi

ctio

n 3:

usi

ng in

dust

ry d

eman

d, in

dust

ry w

ages

by

year

387.

40

110.

80

45.5

n.a.

C. A

ll go

ods

Pre

dict

ion

1: u

sing

ave

rage

dem

and,

ave

rage

wag

es in

200

041

1.5

115.

521

6.0

30.8

Pre

dict

ion

2: u

sing

indu

stry

dem

and,

indu

stry

wag

es in

200

041

1.5

115.

515

5.8

22.7

27.9

26.3

Pre

dict

ion

3: u

sing

indu

stry

dem

and,

indu

stry

wag

es b

y ye

ar

411.

5

115.

5

61.0

3.

3

72

89.3

Sou

rce:

Aut

hors

cal

cula

tion

s as

exp

lain

ed in

the

text

.N

ote:

n.a

. � n

ot a

vaila

ble.


10. For convenience, we omit the petroleum and mining sectors in fi gure 5.3 as well as several other smaller sectors.

are graphed in fi gure 5.3, along with the industry wages in 2000.10 Sectors with the greatest increase in demand include a few tradable industries, like instruments and office machinery and electronic and telecommunication equipment, but many more nontraded goods: real estate, restaurants, scien-tifi c research, education, public administration, health and social services, and so on. At the far right of the fi gure, sectors like textile, wearing apparel, food products, furniture, and agriculture all have negative growth in real demand over 1997 to 2002. We fi nd it quite remarkable that the rapidly growing Chinese economy did not generate more domestic demand for its own tradable goods over this period! Domestic demand should be treated as a viable alternative to exports as a source of employment growth but did not function in that way, presumably because the income gains in China did not lead to a commensurate rise in consumption. Aziz and Cui (2007) argue that one reason for this outcome is that household income did not rise by as much as GDP.

That estimate for rising employment due to nontraded goods can be com-bined with the fall in employment in tradable goods to obtain a total implied change in employment of 61 million workers (prediction 3, panel C). That is our fi nal estimate for 1997 to 2002. In principle, this estimate of 61 � 3.3 � 64.3 million jobs added over 1997 to 2002, from both domestic demand and exports, should equal the actual gain in employment of 39 million jobs. The

Fig. 5.3 Growth in domestic demand, 1997–2002, and industry wages, 2000


11. Note that in fi gure 5.3, the industry with tradable- good industry with the highest percent-age increase in domestic demand is instruments and office machinery, followed by electronic and telecommunication equipment. The latter industry has among the highest wage of any tradable industry and also shows the highest percentage increase in exports (both for ordinary and processing exports).

discrepancy between these numbers (25 million) can be due to multiple causes: we have not been able to distinguish domestic demand by fi rm ownership or province; we have used a fi xed 2000 IO table; and the wage data we use is not as detailed as we would like. But we feel that even if these improvement were made to our calculations, the overall message of table 5.7 would not change: the vast majority of job growth over 1997 to 2002 is due to the increase in demand for nontraded goods, especially the construction sector. The main reason that employment has grown as much as it has in China over 1997 to 2002 is due to the increase in domestic demand for nontradable goods!

Furthermore, it is important to recognize that our fi nal estimate of 61 million jobs gained over 1997 to 2002, from prediction 3, is vastly better than our initial calculation of 216 million jobs (prediction 1, panel C). Comparing these two numbers, we see that the initial calculation is reduced by 72 percent due to the adjustments we have made. That is nearly the same adjustment (75 percent) that we found in the previous section for processing trade but larger than the adjustment (45 percent) that we found for ordinary exports. It is noteworthy that a downward adjustment of 45 percent is shown in table 5.7 for the nontradable sector, where the employment gains were reduced from 203 million in our initial calculation to 61 million (prediction 3). The fact that total employment generated from domestic demand is revised down-ward by nearly 75 percent refl ects the very weak growth in demand for trad-able goods, leading to negative employment gains once we account for the industry composition of demand and wage increases over time. In other words, the Denison effect operates very strongly in the pattern of domestic demand for tradable goods, as we have already seen for exports.11

5.7 Implied Growth in Employment Once Again

Let us now summarize what we have learned from the last three sections and return to the calculations of employment growth. In table 5.8, we show again the static employment coefficients for 2000 (our calculations) and 2002 (from Lau et al. 2006b,c). We found in section 5.2 that those coefficients vastly over-state the actual change in employment over 1997 to 2002 or 2000 to 2005. But by using improved calculations, we were able to reduce the predicted employ-ment growth. Our fi nal calculations showed that the employment growth for ordinary exports was 55 percent lower than obtained from the static employment coefficients, while that employment growth from processing exports and domestic demand were 75 percent lower (and possibly more). We apply those rules of thumb to the initial static employment coefficients to obtain revised employment coefficients, as shown in table 5.8.


12. Note that the predicted employment gains in table 5.9 are not exactly the same as the fi nal row of table 5.7 because in table 5.9, we are using the rules of thumb shown in table 5.8 to reduce the static employment coefficients, that is, the coefficient for ordinary exports is reduced by 55 percent, and the coefficients for processing exports and domestic demand are reduced by 75 percent. Those rules of thumb are broadly consistent but not identical to the calculations in the fi nal row of table 5.7.

For example, instead of the initial calculations for the 2000 IO table, we now predict that $1,000 in ordinary exports generates 0.44 0.45 � 0.20 person- years of employment, while $1,000 in processing exports or domestic demand generates 0.13 0.25 � 0.03 and 0.53 0.25 � 0.13 person- years, respectively. For 2002, we now predict that $1,000 in ordinary exports gen-erates 0.36 0.45 � 0.16 person- years of employment, while $1,000 pro-cessing exports of domestic demand generates 0.11 0.25 � 0.03 and 0.44 0.25 � 0.11 person- years, respectively. These estimates are upper bounds because we obtained lower employment impacts in some calculations, but we shall use these adjustments as conservative.

We use the revised employment coefficients in table 5.8 to recalculate the employment gains for both periods, as shown in table 5.9. For 1997 to 2002, we fi nd that the growth in domestic demand (for nontradable goods, in particular), leads to an increase in employment of 57.8 million workers. In addition, the growth in exports (for ordinary exports, especially), leads to an increase in employment of 12.2 million workers, or about 2.5 million workers per year. Summing over domestic demand and exports, we predict employ-ment gains of 70 million from 1997 to 2002, as compared to the actual employment increase of 39 million.12 So our prediction is nearly twice as big as the actual gain, but that is a great improvement over our initial calculation (table 5.4), where the predicted employment gain was 216 million—more than fi ve times greater than the actual increase! The gap between our revised

Table 5.8 Revised employment coefficients (implied employment increase per $1,000 of exports or domestic demand in person- years)

Source Ordinary exports Processing exports Domestic demand

Author estimates from 2000 IOa 0.444 0.130 0.562

Revised estimates for 2000 IOb 0.444 0.45 � 0.20 0.130 0.25 � 0.03 0.562 0.25 � 0.14

Lau et al. from 2002 IOc 0.363 0.111 0.492Revised estimates for 2000

IOb 0.363 0.45 � 0.16 0.111 0.25 � 0.03 0.492 0.25 � 0.12

Note: IO � input- output.aFrom table 5.3.bRevised as explained in the text and shown in the table.cLau et al. (2006c, table 4), Chen et al. (2008, table 1), and from table 5.3.

Tab

le 5

.9

Impl

ied

Chi

nese

em

ploy

men

t fro

m d

omes

tic

dem

and

and

expo

rts

(usi

ng re

vise

d em

ploy

men

t coe

ffici

ents

)

Gro

wth

in d

eman

d (U

.S.$

bi

llion

s, 2

000)

Impl

ied

empl

oym

ent i

ncre

asea

(mill

ions

of

pers

ons)

Act

ual e

mpl

oym

ent

incr

ease

b (m

illio

ns

of p

erso

ns)

Sour

ce

Per

iod

D

omes

tic

dem

and

E

xpor

ts

Dom

esti

c de

man

d

Exp

orts

To

tal

Aut

hor

esti

mat

es fr

om 2

000

IO19

97–2

002

411.

511

5.5

57.8

12.2

70.0

39.2

Lau

et a

l. (2

006c

) fro

m 2

002

IO

2000

–200

5

735.

0

422.

7

90.4

37

.9

128.

3

37.4

Not

e: I

O �

inpu

t- ou

tput

.a U

ses

the

revi

sed

empl

oym

ent c

oeffi

cien

ts fr

om ta

ble

5.8

and

mul

tipl

ies

thes

e by

the

real

gro

wth

in d

omes

tic

dem

and,

ord

inar

y ex

port

s, a

nd p

roce

ssin

g ex

port

s.b A

ctua

l em

ploy

men

t inc

reas

e co

mes

from

the

Chi

na S

tati

stic

al Y

earb

ook

of e

ach

year

(see

tabl

e 5.

1).


employment gain over 1997 to 2002 and the actual is due to the fall in the labor coefficients B j

Lit from the IO table, refl ecting technological progress and capital accumulation.

In later period, 2000 to 2005, the growth in domestic demand and exports are both stronger. We again use the revised employment coefficient from table 5.8 for 2002 and multiply those by the real changes in domestic demand and exports. We fi nd that the growth in domestic demand (especially invest-ment) leads to an increase in employment of 90.4 million workers. In addi-tion, the growth in exports adds employment of another 38 million work-ers. By coincidence, the predicted employment impact of exports is nearly exactly equal to the actual rise in employment of 37 million workers, or 7.5 million per year.

However, the role of domestic demand over 2000 to 2005, which added 90.4 million to employment is more than twice as large as the role of exports. Based on that evidence, we could not refute the claim that domestic demand is responsible for the employment increase. Whether we want to claim that it is domestic demand or exports that are responsible is really just an exercise in semantics, however: the fact is that both have played an important role in stimulating employment growth, and the sum of them (128.3 million) is still considerably larger than the actual employment gains (37.4 million) over this period. Again, we would attribute the gap between the predicted and actual employment gains as due to technological progress and capital accumula-tion, as well as illustrating the limits of how far we can push our calcula-tions from the IO table. We have made a substantial improvement over the initial calculations, whose predictions were off by an order of magnitude, but still have not obtained a precise accounting of the causes of employ-ment growth.

5.8 Conclusions

Dooley et al. (2003, 2004a,b,c) argue that the current systems of current account imbalances is sustainable so long as China is willing to absorb the Treasury bills used to fi nance the U.S. defi cits. And that willingness is tied to its desire to move workers from unproductive rural employment into urban, manufacturing jobs. These authors suggest that China needs to reemploy some 200 million persons from the countryside, or 10 to 12 million persons per year in the urban areas, and that growth in exports will explain about 30 percent if these employment gains.

We have evaluated this hypothesis by using calculations on the employment impact of exports and domestic demand from Chinese IO tables. We have started with the calculations of Chen et al. (2004) and Lau et al. (2006b,c) for 1995 and 2002 and added our own calculation for the 2000 IO table. The static employment coefficients obtained from these tables summarize the amount of employment generated by $1,000 in exports or domestic demand


for one year. By construction, these static employment coefficients are con-sistent with the full- employment condition for the economy. But the static employment coefficients do a very poor job at predicting the future growth in employment from the future growth in exports or domestic demand. We have shown that the errors involved in this forward- looking forecast are enormous, which means that the static employment coefficients are highly unreliable for that purpose.

To improve on that situation, we have proposed adjustments to the static employment coefficients. These adjustments take into account the future growth in export and domestic industries, which may be quite different from their former growth, as well as rising wages over time. The adjustments partially close the gap between predicted and actual employment growth, even when using an IO table for a single year. Using the revised employment coefficients, we fi nd that export growth over 1997 to 2002 explains at most one- third of the total employment growth in the economy (2.5 out of 7.5 to 8 million workers per year). For 2000 to 2005, however, export growth was faster and, in principle, can explain the entire employment growth of 7.5 million workers per year. However, the rise in domestic demand—especially for investment—generated employment gains that are more than two times larger than those for exports, which confi rms the relative importance of exports as compared to domestic demand suggested by Dooley et al. (2004a). The same amount of employment is reduced by productivity growth in the economy, so the net gain is back to 7.5 million workers per year, somewhat less than the goal put forth by Dooley et al. (2003, 2004a,b,c).

The other key fi nding is that over 1997 to 2002, the rise in domestic demand was nearly entirely in the nontradable sector: predicted employment for tradable goods actually fell. This is very surprising but refl ects the shift in expenditure in China toward construction projects as well as nontradable consumer goods. We do not have the detailed data to evaluate whether the same shift occurred during 2000 to 2005, but from the aggregate GDP data, there has been substantially faster growth in investment instead of in private and public consumption C � G. So we speculate that domestic demand for tradable goods continues to lag, despite the newspaper reports of rising consumer expenditures; this view is also put forth by Aziz and Cui (2007), who point to the slow growth in household income as an explanation.

The importance of this fi nding is that China could certainly turn toward domestic demand instead of export (and consumer expenditures, in par-ticular) as an engine to stimulate employment. The transition from export- led growth to domestic demand would undoubtedly rely on many economic and policy actions that are now only beginning: a real appreciation as the prices of nontradable goods begin to rise, shifting domestic demand toward both imports and exportable goods; accompanied by some nominal appre-ciation of the yuan; fi scal policies that allow for greater security of income in old age, allowing higher expenditures today; reform of the banking sector;


and so on. We believe that it is these features—and not the reliance on export- led growth—that should determine the future path of the govern-ment and trade accounts in China and ultimately restore greater balance to these accounts.

Appendix

Chinese Input- Output Table

The structure of China’s extended IO table separates domestic input from ordinary and processing imported inputs. The direct input requirement coefficient matrix is presented in table 5A.1:

• ADD, ADO, ADP are (n n) matrixes of direct input requirement of domestic products for one unit of domestic product, ordinary export, and processing export, respectively.

• AMD and AOO are the (n n) matrixes of direct input requirement coefficients of ordinary import for one unit domestic production and ordinary exports.

• APP is the (n n) matrix of direct processing import requirement coefficient of producing one unit processing export.

• AVD, AV

O, and APV are each an 1 n vector of direct value added caused

by one dollar of sector j’s production in domestic products, ordinary export, or processing exports.

• ALD, AL

O, APL are each an (1 n) vector of direct labor demand generated

by one dollar production of domestic products, ordinary export, or processing exports.

• EO and EP are each an n 1 vector of ordinary export and processing export, respectively.

Total Value Added (VA) Coefficient Matrix

To calculate the total economy value added, we must consider the linkages between sectors. When one unit domestic product is produced, it generates a fi rst round of value added, which is the direct value added AV

D. However, in order to produce this unit of domestic product, intermediate inputs must be used. The production of these intermediate inputs hence creates the second round of value added, which is named indirect value added (AV

D • ADD). This process of creating indirect value added can continue on and on, as interme-diate inputs are needed to produce other intermediate inputs. Therefore, the total domestic VA induced by a unit domestic production is the sum of fi rst round direct domestic VA and all the indirect domestic VA. Hence, we derive the total domestic VA coefficient (BV

D) aroused by domestic production as:


(A1) BVD � AV

D � AVD • ADD � AV

D • ADD • ADD � AVD • (ADD)3

� . . . � AVD • (I � ADD)�1

Similarly, producing one unit of ordinary or processing export products also requires domestic made intermediate goods, which in turn generates many rounds of VA from these domestic intermediate inputs. We thus have:

(A2) BPV � AP

V � BVD • ADP.

(A3) BVO � AV

O � BVD • ADO,

where BiV represents the total VA coefficient vector for production i, for i �

D (domestic), O (ordinary), and P (processing) respectively.For the same reason, total import content caused by Domestic Production

and exports are defi ned using i as a (1 n) vector of one’s:

(A4) BDM � i • AMD • (I � ADD)�1.

(A5) BOM � i • AOO � i • AMD • (I � ADD)�1 • ADO.

(A6) BPM � i • APP � i • AMD • (I � ADD)�1.

This is conceptually similar to the vertical specialization (VS) as in Hum-mels, Ishii, and Yi (2001) and Dean, Fung, and Wang (2007).

We estimate the total value added B using equations (A1) to (A6). The results are reported in table 5A.2.

Other Data

In table 5A.3, we show the allocation of value added to labor and capital, along with the share of value added within the sum of value added plus

Table 5A.1 Input- output table: direct input requirement coefficient matrix

Domestic product

Ordinary export

Process export Subtotal C � I � G Export

Total output

Domestic intermediate input ADD ADO ADP XD

Ordinary import input AMD AOO 0

Process import input 0 0 APP

Value added ADV AO

V APV

Labor ADL AO

L APL

Total input XD EO EP

Tab

le 5

A.2

To

tal v

alue

add

ed a

nd im

port

con

tent

, 200

0

Inpu

t- ou

tput

indu

stri

es

BV

D

BV

O

BV

P

BM

D

BM

O

BM

P

1. A

gric

ultu

re0.

969

0.89

60.

625

0.03

10.

104

0.37

5 2

. Coa

l min

ing

and

proc

essi

ng0.

945

0.81

10

0.05

50.

189

0 3

. Cru

de p

etro

leum

and

nat

ural

gas

pro

duct

s0.

957

0.81

40.

762

0.04

30.

186

0.23

8 4

. Met

al o

re m

inin

g0.

908

0.62

30.

370

0.09

20.

377

0.63

0 5

. Non

- fer

rous

min

eral

min

ing

0.94

40.

772

0.44

30.

056

0.22

80.

557

6. M

anuf

actu

re o

f fo

od p

rodu

cts

and

toba

cco

proc

essi

ng0.

965

0.90

90.

474

0.03

50.

091

0.52

6 7

. Tex

tile

goo

ds0.

956

0.89

90.

256

0.04

40.

101

0.74

4 8

. Wea

ring

app

arel

, lea

ther

, fur

s, d

own,

and

rel

ated

pro

duct

s0.

958

0.90

90.

171

0.04

20.

091

0.82

9 9

. Saw

mill

s an

d fu

rnit

ure

0.91

50.

674

0.22

50.

085

0.32

60.

775

10. P

aper

and

pro

duct

s, p

rint

ing,

and

rec

ord

med

ium

rep

rodu

ctio

n0.

928

0.76

00.

335

0.07

20.

240

0.66

511

. Pet

role

um p

roce

ssin

g an

d co

king

0.86

50.

268

0.34

30.

135

0.73

20.

657

12. C

hem

ical

s0.

923

0.66

40.

345

0.07

70.

336

0.65

513

. Non

met

al m

iner

al p

rodu

cts

0.92

60.

737

0.35

20.

074

0.26

30.

648

14. M

etal

s sm

elti

ng a

nd p

ress

ing

0.90

10.

635

0.40

40.

099

0.36

50.

596

15. M

etal

pro

duct

s0.

901

0.65

50.

404

0.09

90.

345

0.59

616

. Mac

hine

ry a

nd e

quip

men

t0.

890

0.59

10.

347

0.11

00.

409

0.65

317

. Tra

nspo

rt e

quip

men

t0.

895

0.64

70.

311

0.10

50.

353

0.68

918

. Ele

ctri

c eq

uipm

ent a

nd m

achi

nery

0.89

90.

680

0.17

40.

101

0.32

00.

826

19. E

lect

roni

c an

d te

leco

mm

unic

atio

n eq

uipm

ent

0.85

50.

702

0.18

40.

145

0.29

80.

816

20. I

nstr

umen

ts, m

eter

s, c

ultu

ral a

nd o

ffice

mac

hine

ry0.

857

0.55

00.

191

0.14

30.

450

0.80

921

. Mai

nten

ance

and

rep

air

of m

achi

nery

and

equ

ipm

ent

0.90

70

00.

093

00

22. O

ther

man

ufac

turi

ng p

rodu

cts

0.92

90.

767

0.38

50.

071

0.23

30.

615

23. S

crap

and

was

te1

00

00

024

. Ele

ctri

city

, ste

am, a

nd h

ot w

ater

pro

duct

ion

and

supp

ly0.

930

00

0.07

00

025

. Gas

pro

duct

ion

and

supp

ly0.

921

00

0.07

90

026

. Wat

er p

rodu

ctio

n an

d su

pply

0.95

40

00.

046

00

27. C

onst

ruct

ion

0.91

60.

723

00.

084

0.27

70

28. T

rans

port

and

war

ehou

sing

0.94

40.

806

0.71

70.

056

0.19

40.

283

29. P

ost a

nd te

leco

mm

unic

atio

n0.

941

0.85

00.

388

0.05

90.

150

0.61

230

. Who

lesa

le a

nd r

etai

l tra

de0.

949

0.83

50.

672

0.05

10.

165

0.32

831

. Eat

ing

and

drin

king

pla

ces

0.96

70.

921

0.27

40.

033

0.07

90.

726

32. P

asse

nger

tran

spor

t0.

928

0.76

60.

633

0.07

20.

234

0.36

733

. Fin

ance

and

insu

ranc

e0.

974

0.90

70.

875

0.02

60.

093

0.12

534

. Rea

l est

ate

0.96

80

00.

032

00

35. S

ocia

l ser

vice

s0.

929

0.76

90.

579

0.07

10.

231

0.42

136

. Hea

lth

serv

ices

, spo

rts,

and

soc

ial w

elfa

re0.

927

0.74

10

0.07

30.

259

037

. Edu

cati

on, c

ultu

re a

nd a

rts,

rad

io, fi

lm, a

nd te

levi

sion

0.95

70.

871

0.75

50.

043

0.12

90.

245

38. S

cien

tifi c

res

earc

h0.

893

00

0.10

70

039

. Gen

eral

tech

nica

l ser

vice

s0.

951

0.82

40.

622

0.04

90.

176

0.37

840

. Pub

lic a

dmin

istr

atio

n an

d ot

her

sect

ors

0.

944

0.

807

0

0.

056

0.

193

0


imports used for each type of production: domestic production, ordinary exports, and processing exports. For each type of production, about one- half of value added goes toward compensating labor, with the remainder divided between capital income (one- third) and taxes on production (one- sixth). The amount of value added differs a great deal across type of produc-tion, however: it is 94 percent of the sum of value added plus imports used in domestic production, 62 percent for ordinary exports, and 20 percent for processing exports.

We have also confi rmed that the employment levels in table 5.1 are consis-tent with the IO table itself, as described in table 5A.4. In the fi rst column, we list the economywide compensation to labor from the various years of the IO tables and, in the next columns, the real agricultural and manufacturing wages (in US$, 2000). China employs one- half of its workers in agricul-ture and one- half in manufacturing, so we take the simple average of these two wages to obtain the average wage, which is $842 in 2000, for example. Dividing the labor compensation from the IO table by the average wage, we obtain employment of 716.5 million persons in 2000, which is very close to the 720.5 million persons reported in table 5.1.

For years before and after 2000, however, there is an inconsistency between the actual employment fi gures reported by the China Statistical Yearbook, in the last column of table 5.4, and the implied employment obtained by dividing total compensation from the IO tables by average wages from the China Statistical Yearbook, in the second- to- last column. Implied employ-ment even falls over 1997 to 2002, which does not seem believable. The problem appears to be an inconsistency between the wage series we use (from the China Statistical Yearbook) and the wages that are implicit in the IO

Table 5A.3 Division of value added, 2000 and 2002

From 2000 input- output table 2002

Domestic

production Ordinary exports

Processing exports

Combined production

Combined production

Value added/(Value added � imports) 0.94 0.62 0.20 0.36 n.a.

Compensation of employees/Value added 0.54 0.50 0.45 0.54 0.48

Net taxes on production/Value added 0.14 0.16 0.18 0.15 0.14

Gross capital income/Value added 0.31 0.34 0.37 0.31 0.37

Notes: n.a. � not available. Figures reported here are only for the direct use of labor and imports in each type of production and do not take into account the indirect usage through domestic intermediate in-puts.

Tab

le 5

A.4

W

ages

and

em

ploy

men

t

Yea

r

Com

pens

atio

n fr

om

IO ta

ble

(U.S

.$ m

illio

ns, 2

000)

Agr

icul

ture

se

ctor

wag

e (U

.S.$

, 200

0)

Man

ufac

turi

ng

sect

or w

age

(U.S

.$, 2

000)

Rea

l wag

ea (a

vera

ge;

U.S

.$, 2

000)

Impl

ied

empl

oym

entb

(mill

ions

of

pers

ons)

Act

ual e

mpl

oym

entc

(mill

ions

of

pers

ons)

1995

334,

000

476.

669

9.4

520

641.

968

0.7

1997

501,

101

557.

976

7.8

618

811.

069

8.2

2000

603,

003

626.

210

57.0

842

716.

572

0.9

2002

712,

224

740.

012

72.4

1,00

670

8.0

737.

420

05n.

a.89

4.1

1695

.51,

468

758.

3

Usi

ng r

evis

ed w

age

data

d

1997

501,

101

602.

582

9.2

715.

970

0.0

698.

220

02

712,

224

71

0.4

12

21.5

96

6.0

73

7.3

73

7.4

Not

e: I

O �

inpu

t- ou

tput

; n.a

. � n

ot a

vaila

ble.

a Ave

rage

wag

e is

the

sim

ple

aver

age

of m

anuf

actu

ring

and

agr

icul

ture

sec

tors

. Sou

rce

for

wag

e da

ta is

the

Chi

na S

tati

stic

al Y

earb

ook,

200

6.b I

mpl

ied

empl

oym

ent �

rea

l com

pens

atio

n fr

om I

O ta

ble/

real

ave

rage

wag

e.c A

ctua

l em

ploy

men

t dat

a co

me

from

Chi

na S

tati

stic

al Y

earb

ook

of e

ach

year

.d T

he r

evis

ed w

age

data

mul

tipl

ies

1997

wag

es b

y 1.

08, a

nd m

ulti

plie

s 20

02 w

ages

by

0.96

so

that

the

impl

ied

empl

oym

ent i

s ro

ughl

y eq

ual t

o ac

tual

em

ploy

men

t.


13. For 2000, when we have the most complete IO table available, it lists both labor compen-sation and employment at the end of the year. So the wages being used in the IO table can be computed, and they are highly consistent with both the wages and actual employment fi gures used in table 5.4 for 2000. For 1997 and 2002, however, the IO table is less complete and, in particular, does not list employment so that implied wages cannot be computed.

tables, at least in 1997 and 2002.13 It is essential that the implied employment from the IO table in each year equal actual employment in the economy. To achieve this, we infl ate the 1997 wages from the China Statistical Yearbook by 8 percent and defl ate the 2002 wages by 4 percent, obtaining the revised wages reported in the bottom of table 5.4. Those adjusted wages lead to implied employment from the IO tables that is roughly equal to that reported by the China Statistical Yearbook. We will continue to use this simple adjust-ment to 1997 and 2002 wages in all our calculations.

References

Aziz, Jahangir, and Li Cui. 2007. Explaining China’s low consumption: The neglected role of household income. IMF Working Paper no. 07/181. Washington, DC: International Monetary Fund.

Bernstein, Jeffrey R., and David E. Weinstein. 2002. Do endowments predict the location of production? Evidence from national and international data. Journal of International Economics 56 (1): 55–76.

Chen, Xikang, Leonard K. Cheng, K. C. Fung, and Lawrence J. Lau. 2004. The estimation of domestic value added and employment induced by exports: An application to Chinese exports to the United States. Paper presented at the AEA conference, Boston.

Chen, Xikang, Leonard K. Cheng, K. C. Fung, Lawrence J. Lau, Yung Wing Sung, C. Yang, K. Zhu, J. Pei, and Z. Tan. 2008. Domestic value added and employment generated by Chinese exports: A quantitative estimation. FREIT Working Paper no. 7. http://eitt.org/WorkingPapers/Papers/TradePolicyGeneral/FREIT007.pdf.

Dean, Judith M., K. C. Fung, and Zhi Wang. 2007. Measuring the vertical specializa-tion in Chinese trade. U.S. International Trade Commission, Working Paper.

Denison, Edward. 1967. Why growth rates differ. Washington, DC: Brookings Insti-tution.

———. 1980. Accounting for slower growth: The United States in the 1970s. Wash-ington, DC: Brookings Institution.

Dooley, Michael P., David Folkerts- Landau, and Peter Garber. 2003. An essay on the revived Bretton Woods system, NBER Working Paper no. 9971. Cambridge, MA: National Bureau of Economic Research.

———. 2004a. Direct investment, rising real wages and the absorption of excess labor in the periphery. NBER Working Paper no. 10626. Cambridge, MA: Na-tional Bureau of Economic Research. Also published in G7 Current Account Imbalances: Sustainability and Adjustment, Richard H. Clarida, ed., 103–27. Chi-cago: The National Bureau of Economic Research, 2007.

———. 2004b. The revived Bretton Woods system: The effects of periphery interven-tion and reserve management on interest rates and exchange rates in center coun-


tries. NBER Working Paper no. 10331. Cambridge, MA: National Bureau of Economic Research.

———. 2004c. The U.S. current account defi cit and economic development: Col-lateral for a total return swap. NBER Working Paper no. 10727. Cambridge, MA: National Bureau of Economic Research.

———. 2005. International Financial Stability: Asia, Interest Rates, and the Dollar. Deutsche Bank. http://people.ucsc.edu/~mpd/InternationalFinancialStability_update.pdf.

Felipe, Jesus, and Rana Hasan. 2006a. The challenge of job creation in Asia. ERD Policy Brief no. 44. Manila, The Philippines: Asian Development Bank. http://www.asiandevbank.org/Documents/EDRC/Policy_Briefs/PB044.pdf.

———. 2006b. Labor markets in a globalizing world. In Labor markets in Asia: Issues and perspectives, ed. Jesus Felipe and Rana Hasan, 63–142. London: Palgrave Macmillan.

Hummels, David, Jun Ishii, and Kei- Mu Yi. 2001. The nature and growth of vertical specialization in world trade. Journal of International Economics 54:75–96.

Lau, Lawrence J., Xikang Chen, Leonard K. Cheng, K. C. Fung, Yun- Wing Sung, Cuihong Yang, Kunfu Zhu, Zhipeng Tang, and Jiansuo Pei. 2006a. Estimates of U.S.- China trade balances in terms of domestic value added. Stanford Center for International Development Working Paper no. 295. Stanford, CA: Stanford Uni-versity.

———. 2006b. The estimation of domestic value added and employment generated by U.S.- China trade. Institute of Economics Working Paper no. 2. Hong Kong: Chinese University of Hong Kong.

———. 2006c. Estimation of domestic value added and employment induced by China’s exports. Unpublished Manuscript.

Nordhaus, William. 2002. Productivity growth and the new economy. Brookings Papers on Economic Activity, Issue no. 2:211–65. Washington, DC: Brookings Institution.

Wei, Shang- Jin. 2007. Comment on Direct Investment, rising real wages, and the absorption of excess labor in the periphery. In G7 Current Account Imbalances: Sustainability and Adjustment, ed. Richard Clarida, 128–32. chicago: National Bureau of Economic Research

Comment Michael Dooley

This chapter provides a careful evaluation of the contribution of export growth to growth in manufacturing employment in China in recent years. Feenstra and Hong are generous in citing my work with Folkerts- Landau and Garber as providing an incentive to evaluate the role of export industries in absorbing some 250 million underemployed workers over the next decade or so. They fi nd that our back- of- the- envelope calculation that exports have generated one- third of the growth in employment is roughly consistent with their estimates generated from input- output data adjusted for productivity

Michael Dooley is a professor of economics at the University of California, Santa Cruz, and a research associate of the National Bureau for Economic Research.

China's Growing Role in World Trade - Clark University NBER Trade Volume (2010) China's...exports....

Documents

Transcript of China's Growing Role in World Trade - Clark University NBER Trade Volume (2010) China's...exports....